One thing missing from the hypermedia section: links do not have to live in the payload. RFC 8288 lets you carry the same relations in a Link header, and for some clients that is the only place they look. I advertise a markdown twin of a page as Link: </llms.txt>; rel="describedby"; type="text/markdown" for exactly that reason, no body parsing needed.
Conditional requests feel like the other omission. ETag and 304 are part of the same "use HTTP as intended" argument, and they change client behaviour more often than whether someone hardcoded a route.
> REST (Representational State Transfer) is a software architectural style that was created to describe the design and guide the development of the architecture for the World Wide Web. REST defines a set of constraints for how the architecture of a distributed, Internet-scale hypermedia system, such as the Web, should behave.
And isn't it always misused, mis-represented basically all the time? Often confused with things that are basically just JSON API(s)?
Any descriptive term where people can't agree on what it means is completely useless. So there's no point trying to argue "REST means this" - the fact that so many people have done that already makes it pointless.
The example given in this section... is there actually a commonly used REST api that uses that particular key/structure? I wouldn't be _surprised_ if I came across that in some random API's documentation, but I hardly think it's the 'usual' thing to do.
The article later claims that that style is from https://stateless.co/hal_specification.html , which I think is a consultancy company whose major partners are all casinos (am I totally off on this?). The article also references (https://github.com/kevinswiber/siren which adds detail re: http methods embedded in the _links structure, but I'm not sure why that repo should be considered authoritative. 1300 stars for a just a spec is actually not bad though I think, so maybe.
Overall this article seems like slop produces from some kind of ai crawler that managed to dig up a couple of random links.
REST seems to bring out people who like to define specs and correct others on following the spec or not. There are many other things in tech like this.
To be honest, I just kind of don't understand the point. Could my API be better following some of these rules the author presented? Probably. But by how much? Is it worth it? When someone else on my team disagrees do we go back and argue about the spec again?
After 12+ years I'm just not sure anymore. If I'm using REST incorrectly (according to this article or others), but getting good results and everyone on the team is aligned, does it matter? I'm willing to have my mind changed this, maybe I'm missing something important
> REST seems to bring out people who like to define specs and correct others on following the spec or not. There are many other things in tech like this.
Yes, and this blog post is exactly yet another example. Some of the blogger's points are blatant violations of REST, such as the comments on URIs and api versioning. Yet, the blogger elevates their own personal misconceptions to the status of one true RESTful way.
Ultimately it's all mostly irrelevant. All it matters is that the contract is specified and preserved and clients can consume services through a specific interface. Furthermore the pervasive use of openapi specs already crystalize the RPC approach to interfaces.
I try not to blame the 'Roy Fielding is right' camp on this. I also try not to blame the "terminology doesn't matter' camp.
There are groups of people using the same terms for different things, and getting into arguments where there could be agreement.
There's a diverse set of people on the spectrum between formally trained and self taught people in software development. Terms can get borrowed from places like academic research and get (mis)applied before they are understood.
The REST in industry software is essentially JSON RPC (no, not that one) mapped to HTTP methods. That's a mouthful, so people just say RESTful, despite it not quite being RESTful according to the original notion of REST.
Yes, industry-RESTful isn't necessarily RESTful according to original Roy Fielding definition, but trying to redefine a widely used industry term is unlikely to win one friends at dinner parties.
whole thing about RESTful api - majority of people never understood it - few people like the htmx crowd tried to bring it up again to teach people the right way.
but people who didn't understand rest some of them migrated to graphql or rpc.
Many years ago as a young developer I read Fielding's dissertation, Fowler's writings on the Richardson Maturity Model and even a book (don't remember the title) on the topic. Then I spent the next dozen years creating, reading, updating and deleting JSON APIs.
Problem Details (RFC 9457) are useful and have made an occasional appearance, but actual hypermedia driven APIs with custom media types seem to be non-existent outside of example blog posts. It's much more common to instead see APIs overload lesser-used HTTP status codes with their own domain-specific meanings. But if I'm being honest, this sort of protocol crime almost never caused an actual problem in practice.
The advocacy for "true REST" is missing a compelling argument and useful example of why it would actually be an improvement. Despite its prominence, for over 20 years almost nobody has used this approach. Why is that?
I do still have an aesthetic appreciation for "true REST", but in practice I find the fact that nobody uses it to be mainly useful in ending bikeshedding discussions. Once you understand your API isn't RESTful to begin with, there's no longer any point in arguing if 411 or 412 is the more RESTful status code to use when a user attempts to order a product without specifying a quantity. (Spoiler: neither is correct) You simply send a 400 with Problem Details and call it a day. Ironically, this indifference may make your API just a little bit more RESTful.
I had flashbacks to when I first started making web apps. I used to care so very much about this. I haven’t thought about capital REST in years.
Edit: a few minutes after I wrote that I had the thought, what if this is someone else’s arc with AI code. A lot of people seem to care a whole awful lot right now. A decade time horizon is interesting to think about.
REST is an inherently human technology, almost superfluous for API(emphasis on Programming Interface) usage. I guess it does not strictly have to be, especially with our modern LLM's But REST can best be summed up as "ship the human interface required to handle a state with every state". For a strictly programming interface, this interface coupled with the state is inefficient and mostly pointless(the application already knows how to handle the state, it does not have to figure it out) and is better served by a more focused mechanical transfer.
With how narrow this scope ends up being I can understand how most of the tech industry collectively had what appears to be a collective brain fart when Fielding said "hey this web thing is doing something really novel in the field of application design, lets talk about it" and went "herp derp, REST == HTTP, sounds good to me". I mean I understand it, but still find it funny.
I was a REST purist for a while and the purer I got the more unhappy I was. Hypermedia standards like Siren and HAL just made things miserable no matter how hard I tried to make them work.
I think the REST Architectural Pattern is great but I don't use it anymore. Instead I create ROUSED services.
Resource Oriented because the very best thing about REST is the tendency to focus on resources and collections of resources and how you can interact with them sensibly.
Unconstrained because I don't let some specification control how I work.
Sensible because things still need to be internaly consistent and well thought out.
Enterprise because the fact is I'm always making these services for some 'boring' business use case.
Services that provide some specific set of bounded capabilities within the usual messy ecosystem present at a SaaS software company.
I think we owe a lot to Roy Fielding for the potent ideas but as others have argued eloquently, reality sets in and you have to adapt to circumstances.
> You must not store or reuse URI structures in the client. An API could change them without notice, and your client would obviously stop working. Instead, parse the links and follow the relative identifiers (rel).
This part is a complete and total fairy tale. No API consumer wants to operate in a world where the API could "change without notice." In real life, no provider would expect that of their clients, and no client would seriously accept that before every API request, they must first hierarchically rediscover the provider's dynamic url scheme of the moment, in order to traverse down to resource they ultimately want to request.
There has never been any serious commercial or high-traffic API that implemented this. And it's not because "lame corporate tech teams aren't hip to the true meaning of REST" -- it's because "it doesn't actually make sense to do it that way so nobody wants it to work like that."
Every time I look at a "how to do pure REST", it seems like doing REST correctly is building a self-describing database instead of making an API that does a thing.
Here's the thing:
Customers don't care if your application does REST correctly or not. They only care if your application works.
API consumers care about REST only as far as the HTTP API makes sense and follows well-known and well-understood conventions.
I don't think I've ever come across a self-describing API in the wild where I have to constantly self-discover URIs. (And frankly, if someone proposed something like this I'd laugh at them.) It's just needless overengineering.
Anyone working in an organisation where teams are trying to map business logic errors into HTTP status codes?
Wasn't Fieldings original dissertation meant as a formalisation of the web? I.e. it had nothing to do with APIs until later, when that became popular in the mid 2000s?
The amount of nonsense conversations / meetings I have been in over the years around designing "RESTful" APIs, endless discussions, I can no longer keep count. Because, unless you're working on an academic research project, everything in REST becomes a compromise.
At this point I am starting believe that REST has done more harm than good for the industry.
That's actually really correct in the most wrong way possible. I've worked with one too many endpoints that don't care about UPDATE, DELETE, and whatever else beside POST exists, while relying on body data to do the "what department are you looking for" requests.
Yes, I also balked when I originally encountered this pattern at a previous co.
And I think agree with your point, "correct in a wrong way"? Yeah, everything as a POST is wrong, but it is less wrong and more clear to just ignore REST because the assumptions are broken so often it just makes it more complicated.
Just POST everything and make it clear in the URI + bodies what is happening.
31 comments
[ 0.23 ms ] story [ 30.8 ms ] threadConditional requests feel like the other omission. ETag and 304 are part of the same "use HTTP as intended" argument, and they change client behaviour more often than whether someone hardcoded a route.
> REST (Representational State Transfer) is a software architectural style that was created to describe the design and guide the development of the architecture for the World Wide Web. REST defines a set of constraints for how the architecture of a distributed, Internet-scale hypermedia system, such as the Web, should behave.
And isn't it always misused, mis-represented basically all the time? Often confused with things that are basically just JSON API(s)?
https://htmx.org/essays/how-did-rest-come-to-mean-the-opposi...
And here is why REST is probably largely pointless in machine-to-machine APIs:
https://intercoolerjs.org/2016/05/08/hatoeas-is-for-humans.h...
The example given in this section... is there actually a commonly used REST api that uses that particular key/structure? I wouldn't be _surprised_ if I came across that in some random API's documentation, but I hardly think it's the 'usual' thing to do.
The article later claims that that style is from https://stateless.co/hal_specification.html , which I think is a consultancy company whose major partners are all casinos (am I totally off on this?). The article also references (https://github.com/kevinswiber/siren which adds detail re: http methods embedded in the _links structure, but I'm not sure why that repo should be considered authoritative. 1300 stars for a just a spec is actually not bad though I think, so maybe.
Overall this article seems like slop produces from some kind of ai crawler that managed to dig up a couple of random links.
I am glad that it mentions rfc9457 though.
To be honest, I just kind of don't understand the point. Could my API be better following some of these rules the author presented? Probably. But by how much? Is it worth it? When someone else on my team disagrees do we go back and argue about the spec again?
After 12+ years I'm just not sure anymore. If I'm using REST incorrectly (according to this article or others), but getting good results and everyone on the team is aligned, does it matter? I'm willing to have my mind changed this, maybe I'm missing something important
Yes, and this blog post is exactly yet another example. Some of the blogger's points are blatant violations of REST, such as the comments on URIs and api versioning. Yet, the blogger elevates their own personal misconceptions to the status of one true RESTful way.
Ultimately it's all mostly irrelevant. All it matters is that the contract is specified and preserved and clients can consume services through a specific interface. Furthermore the pervasive use of openapi specs already crystalize the RPC approach to interfaces.
There are groups of people using the same terms for different things, and getting into arguments where there could be agreement.
There's a diverse set of people on the spectrum between formally trained and self taught people in software development. Terms can get borrowed from places like academic research and get (mis)applied before they are understood.
The REST in industry software is essentially JSON RPC (no, not that one) mapped to HTTP methods. That's a mouthful, so people just say RESTful, despite it not quite being RESTful according to the original notion of REST.
Yes, industry-RESTful isn't necessarily RESTful according to original Roy Fielding definition, but trying to redefine a widely used industry term is unlikely to win one friends at dinner parties.
but people who didn't understand rest some of them migrated to graphql or rpc.
Problem Details (RFC 9457) are useful and have made an occasional appearance, but actual hypermedia driven APIs with custom media types seem to be non-existent outside of example blog posts. It's much more common to instead see APIs overload lesser-used HTTP status codes with their own domain-specific meanings. But if I'm being honest, this sort of protocol crime almost never caused an actual problem in practice.
The advocacy for "true REST" is missing a compelling argument and useful example of why it would actually be an improvement. Despite its prominence, for over 20 years almost nobody has used this approach. Why is that?
I do still have an aesthetic appreciation for "true REST", but in practice I find the fact that nobody uses it to be mainly useful in ending bikeshedding discussions. Once you understand your API isn't RESTful to begin with, there's no longer any point in arguing if 411 or 412 is the more RESTful status code to use when a user attempts to order a product without specifying a quantity. (Spoiler: neither is correct) You simply send a 400 with Problem Details and call it a day. Ironically, this indifference may make your API just a little bit more RESTful.
Edit: a few minutes after I wrote that I had the thought, what if this is someone else’s arc with AI code. A lot of people seem to care a whole awful lot right now. A decade time horizon is interesting to think about.
https://github.com/NationalBankBelgium/REST-API-Design-Guide...
With how narrow this scope ends up being I can understand how most of the tech industry collectively had what appears to be a collective brain fart when Fielding said "hey this web thing is doing something really novel in the field of application design, lets talk about it" and went "herp derp, REST == HTTP, sounds good to me". I mean I understand it, but still find it funny.
I think the REST Architectural Pattern is great but I don't use it anymore. Instead I create ROUSED services.
Resource Oriented because the very best thing about REST is the tendency to focus on resources and collections of resources and how you can interact with them sensibly.Unconstrained because I don't let some specification control how I work.
Sensible because things still need to be internaly consistent and well thought out.
Enterprise because the fact is I'm always making these services for some 'boring' business use case.
Services that provide some specific set of bounded capabilities within the usual messy ecosystem present at a SaaS software company.
I think we owe a lot to Roy Fielding for the potent ideas but as others have argued eloquently, reality sets in and you have to adapt to circumstances.
This part is a complete and total fairy tale. No API consumer wants to operate in a world where the API could "change without notice." In real life, no provider would expect that of their clients, and no client would seriously accept that before every API request, they must first hierarchically rediscover the provider's dynamic url scheme of the moment, in order to traverse down to resource they ultimately want to request.
There has never been any serious commercial or high-traffic API that implemented this. And it's not because "lame corporate tech teams aren't hip to the true meaning of REST" -- it's because "it doesn't actually make sense to do it that way so nobody wants it to work like that."
Here's the thing:
Customers don't care if your application does REST correctly or not. They only care if your application works.
API consumers care about REST only as far as the HTTP API makes sense and follows well-known and well-understood conventions.
I don't think I've ever come across a self-describing API in the wild where I have to constantly self-discover URIs. (And frankly, if someone proposed something like this I'd laugh at them.) It's just needless overengineering.
Wasn't Fieldings original dissertation meant as a formalisation of the web? I.e. it had nothing to do with APIs until later, when that became popular in the mid 2000s?
The amount of nonsense conversations / meetings I have been in over the years around designing "RESTful" APIs, endless discussions, I can no longer keep count. Because, unless you're working on an academic research project, everything in REST becomes a compromise.
At this point I am starting believe that REST has done more harm than good for the industry.
And I think agree with your point, "correct in a wrong way"? Yeah, everything as a POST is wrong, but it is less wrong and more clear to just ignore REST because the assumptions are broken so often it just makes it more complicated.
Just POST everything and make it clear in the URI + bodies what is happening.